def calculate_x_y(self):
try:
variables = np.load('Games/' + 'MLRxy.npz')
self.x = variables['arr_0'] # init
self.y = variables['arr_1']
self.x_max = self.x.max(axis=0)
self.x_min = self.x.min(axis=0)
except:
obj = MLR()
obj.Extract() # bring the independent variables (X)
obj.Y() # bring in the dependent variable (y)
#total_attr = 13 # add all the 2nd dimensions of each attributes
obj.X() # combine all the independent variables to one
self.x = obj.x # init
self.y = obj.y # init
self.x_max = self.x.max(axis=0)
self.x_min = self.x.min(axis=0)
obj.save()
temp = np.ones(shape=(40, 1))
self.x = np.concatenate(
(self.x, temp), axis=1
) # 40 by 13+1 where last column has all value as ones... to obtain a constant-beta value
return
def result(self, home,
away): #calculate the result for current season here
new = MLR()
new.calculate_points(home, away)
new.count_h2h(home, away)
new.current_standings(home, away)
new.current_season_games(home, away)
new.last_4_games_this_season(home, away)
new.X()
self.X = new.x[0]
self.X = np.matrix(self.X)
# check if x-values lie between the limits of the training set
for i in xrange(13):
if self.X[0, i] > self.x_max[i]:
self.X[0, i] = self.x_max[i]
if self.X[0, i] < self.x_min[i]:
self.X[0, i] = self.x_min[i]
self.X = np.concatenate((self.X, np.matrix('1')), axis=1)
return